redis集群搭建(手动指定master和slave)

浅浅的花香味﹌ 2022-02-17 06:11 779阅读 0赞

接上一篇：[https://blog.csdn.net/u010533511/article/details/89388547][https_blog.csdn.net_u010533511_article_details_89388547]

1、创建3个master

/usr/local/redis_cluster/redis-trib.rb create --replicas 0 192.168.135.23:7001 192.168.135.229:7007 192.168.135.196:7004

\--replicas 1表示为集群的master节点创建1个slave。那么6个实例里，有三个master，有三个是slave。

\--replicas 0不为集群的master节点创建slave。

2、分别指定3个slave到指定的主下面

/usr/local/redis_cluster/redis-trib.rb add-node --slave --master-id d1983a08bade495716392200e2b3efbd09736631 192.168.135.196:7005 192.168.135.23:7001
    
    /usr/local/redis_cluster/redis-trib.rb add-node --slave --master-id 9cd7049834c0f3183b474b0bd3e516b5359f6c67 192.168.135.229:7008 192.168.135.196:7004
    
    /usr/local/redis_cluster/redis-trib.rb add-node --slave --master-id 5d6bc6740e5bc88643377746d140fdf63f812d88 192.168.135.23:7002 192.168.135.229:7007

3、查看集群状态：

[root@localhost /]# cd /usr/local/redis_cluster
    [root@localhost redis_cluster]# ./redis-trib.rb check 192.168.135.23:7001
    >>> Performing Cluster Check (using node 192.168.135.23:7001)
    M: d1983a08bade495716392200e2b3efbd09736631 192.168.135.23:7001
       slots:0-5460 (5461 slots) master
       1 additional replica(s)
    S: 883862d8a29665094995ca75f41c39186979bbc0 192.168.135.229:7008
       slots: (0 slots) slave
       replicates 9cd7049834c0f3183b474b0bd3e516b5359f6c67
    S: d5c04adcad1034d2b1bcb327a86a9519efe1ca04 192.168.135.196:7005
       slots: (0 slots) slave
       replicates d1983a08bade495716392200e2b3efbd09736631
    M: 9cd7049834c0f3183b474b0bd3e516b5359f6c67 192.168.135.196:7004
       slots:10923-16383 (5461 slots) master
       1 additional replica(s)
    S: 2f40bcf6bdde917a39ed67c47e943161b378d952 192.168.135.23:7002
       slots: (0 slots) slave
       replicates 5d6bc6740e5bc88643377746d140fdf63f812d88
    M: 5d6bc6740e5bc88643377746d140fdf63f812d88 192.168.135.229:7007
       slots:5461-10922 (5462 slots) master
       1 additional replica(s)
    [OK] All nodes agree about slots configuration.
    >>> Check for open slots...
    >>> Check slots coverage...
    [OK] All 16384 slots covered.

4、测试一下get 和 set数据

redis-cli -c -p 7001

进入命令窗口，直接 set  test success

直接会根据hash匹配切换到相应的slot的节点上。

redis集群有16383个slot组成，通过分片分布到多个节点上，读写都发生在master节点。

5、关闭某个服务器的所有redis服务查看集群是否正常

比如关掉229的redis服务

[root@localhost /]# ps -ef|grep cluster
    root     21193  6432  0 23:19 pts/1    00:00:00 grep --color=auto cluster
    root     24263     1  0 14:10 ?        00:00:17 ../redis-server 0.0.0.0:7007 [cluster]
    root     26660     1  0 14:32 ?        00:00:16 ../redis-server 0.0.0.0:7008 [cluster]
    [root@localhost /]# kill 24263 26660
    [root@localhost /]# ps -ef|grep cluster
    root     21236  6432  0 23:19 pts/1    00:00:00 grep --color=auto cluster

查看集群是否正常

[root@localhost redis_cluster]# ./redis-cli -c -p 7001
    192.168.135.23:7002> get test
    -> Redirected to slot [12291] located at 192.168.135.196:7004
    "success"
    192.168.135.196:7004> cluster nodes
    d5c04adcad1034d2b1bcb327a86a9519efe1ca04 192.168.135.196:7005@17005 slave d1983a08bade495716392200e2b3efbd09736631 0 1555601184508 7 connected
    d1983a08bade495716392200e2b3efbd09736631 192.168.135.23:7001@17001 master - 0 1555601184120 7 connected 0-5460
    9cd7049834c0f3183b474b0bd3e516b5359f6c67 192.168.135.196:7004@17004 myself,master - 0 1555601182000 9 connected 10923-16383
    883862d8a29665094995ca75f41c39186979bbc0 192.168.135.229:7008@17008 slave,fail 9cd7049834c0f3183b474b0bd3e516b5359f6c67 1555601043264 1555601042000 9 disconnected
    2f40bcf6bdde917a39ed67c47e943161b378d952 192.168.135.23:7002@17002 master - 0 1555601185527 10 connected 5461-10922
    5d6bc6740e5bc88643377746d140fdf63f812d88 192.168.135.229:7007@17007 master,fail - 1555601043264 1555601043181 5 disconnected
    192.168.135.196:7004>

可以看出229挂掉后集群依然正常，实现了高可用，重启229上面的redis服务后会自动加入集群

192.168.135.196:7004> cluster nodes
    d5c04adcad1034d2b1bcb327a86a9519efe1ca04 192.168.135.196:7005@17005 slave d1983a08bade495716392200e2b3efbd09736631 0 1555601482087 7 connected
    d1983a08bade495716392200e2b3efbd09736631 192.168.135.23:7001@17001 master - 0 1555601483301 7 connected 0-5460
    9cd7049834c0f3183b474b0bd3e516b5359f6c67 192.168.135.196:7004@17004 myself,master - 0 1555601479000 9 connected 10923-16383
    883862d8a29665094995ca75f41c39186979bbc0 192.168.135.229:7008@17008 slave 9cd7049834c0f3183b474b0bd3e516b5359f6c67 0 1555601482303 9 connected
    2f40bcf6bdde917a39ed67c47e943161b378d952 192.168.135.23:7002@17002 master - 0 1555601481000 10 connected 5461-10922
    5d6bc6740e5bc88643377746d140fdf63f812d88 192.168.135.229:7007@17007 slave 2f40bcf6bdde917a39ed67c47e943161b378d952 0 1555601482503 10 connected

**但是发现229上的redis重新启动后都变成了slave**，23上面现在是两个master这样的话如果某天23服务器挂掉了，那集群肯定也就挂了，怎么办呢，没找到好的解决办法，于是就把23上面的那个原来的slave的给它kill掉，然后229上面对应的slave就有恢复成了原来的master了，不知道各位大神还有没有什么好的解决这个问题的办法？

[https_blog.csdn.net_u010533511_article_details_89388547]: https://blog.csdn.net/u010533511/article/details/89388547